|
--- |
|
license: mit |
|
base_model: microsoft/phi-2 |
|
tags: |
|
- alignment-handbook |
|
- trl |
|
- sft |
|
- generated_from_trainer |
|
- trl |
|
- sft |
|
- generated_from_trainer |
|
datasets: |
|
- GAIR/lima |
|
model-index: |
|
- name: phi2-lima |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# phi2-lima |
|
|
|
This model is a fine-tuned version of [microsoft/phi-2](https://huggingface.co/microsoft/phi-2) on the GAIR/lima dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 2.5096 |
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
More information needed |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 2e-05 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- distributed_type: multi-GPU |
|
- num_devices: 16 |
|
- total_train_batch_size: 128 |
|
- total_eval_batch_size: 128 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_ratio: 0.1 |
|
- num_epochs: 10.0 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 2.2578 | 1.0 | 6 | 2.3195 | |
|
| 2.1177 | 2.0 | 12 | 2.1448 | |
|
| 2.0262 | 3.0 | 18 | 2.1417 | |
|
| 1.9422 | 4.0 | 24 | 2.2227 | |
|
| 1.7786 | 5.0 | 30 | 2.3327 | |
|
| 1.7224 | 6.0 | 36 | 2.4202 | |
|
| 1.684 | 7.0 | 42 | 2.4698 | |
|
| 1.6434 | 8.0 | 48 | 2.4961 | |
|
| 1.616 | 9.0 | 54 | 2.5094 | |
|
| 1.6183 | 10.0 | 60 | 2.5096 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.38.2 |
|
- Pytorch 2.1.2 |
|
- Datasets 2.14.6 |
|
- Tokenizers 0.15.2 |
|
|
